Hello, Iām Takuro Huang (Wen-Shau Huang).
Motivated software developer with 9 years of experience in New York and Taiwan. Versatile in various areas including mobile development, backend programming, and AI research.
Dedicated team player who thrives in collaborative environments, effectively communicating with cross-functional teams to deliver high-quality software solutions.
Committed to making a positive impact on projects and teams, leveraging technical expertise and problem-solving skills to overcome challenges. Passionate about staying up-to-date with new technologies and continuously expanding knowledge to contribute valuable insights and drive innovation.
Element is a pioneering biometric digital identity company that collaborates with progressive financial institutions, governments, and healthcare organizations to revolutionize service delivery. Co-founded by Yann LeCun, Element holds the distinction of being the first AI company dedicated to digital identity, playing a crucial role in establishing deep learning on mobile devices. As a software developer at Element, I am actively involved in developing an internal license management tool and an innovative data provenance platform. Working closely with both the research and development teams, I contribute to the design and implementation of end-to-end solutions such as a data marketplace, AutoML, and generative models. Leveraging the power of GCP and open-source solutions, we strive to deliver cutting-edge services to our clients.
Artomotion was founded by professors from Feng Chia University and the National Taiwan University of Sport, two professional baseball players, and myself. As a company focused on sports science, our main objective is to develop AI training products for professional players. We are specifically dedicated to building a baseball training system for personalized training and an automated strike zone system for official baseball games, which has already been successfully implemented in the Chinese Professional Baseball League (CPBL). Our automated strike zone system not only detects strikes and balls but also captures videos from multiple angles for in-depth analysis of movement, spin, and axis, among other factors. Given that we are a small startup, I assume the roles of both CEO and CTO within the company. In this capacity, I lead the development team, ensuring the progress and success of our projects.
As the lead mobile developer at Gooroo, an EdTech company providing online and in-person tutoring services in the USA, my primary responsibility was to collaborate with the backend and UI/UX designers. Together, we built native iOS and Android applications for both tutors and students. Leading the team, I ensured that we followed best practices to deliver high-quality and user-friendly experiences.
The AI Research Center at Feng Chia University is one of the largest centers, dedicated to collaborating with enterprises to bring cutting-edge technology into production environments. In my role as a team leader, I led a team in cooperation projects with one of the biggest online retailers in Taiwan. Our focus was to provide an end-to-end production data-driven service using R and Python, aiming to enhance the retailer's commodity designs. To achieve this, we developed a machine learning pipeline capable of analyzing NLP, image, and tabular data, even when dealing with imbalanced labels. Our efforts aimed to leverage advanced technologies and empower students to learn by doing, while delivering practical solutions to improve the retailer's offerings.
At OTG, a hospitality industry innovator, I worked as a software engineer primarily focused on maintaining and delivering features for the iOS app, Concierge. I collaborated with partners like Chase and United Airlines to introduce exclusive membership features. Additionally, I contributed to the development of a Python-based recommendation engine in collaboration with data scientists.
At Gooroo, an NYC-based EdTech company providing in-person tutoring services, I led the development team in creating their first official iOS and Android app for matching tutors with students. I directed the design and development of the cloud-based backend service using Node.js and AWS. Additionally, I implemented a pricing strategy and integrated the Stripe payment method for seamless transactions.
Mobagel was originally a mobile app development company specializing in iOS gaming and applications. However, it has since pivoted to focus on AutoML (Automated Machine Learning). During my time at Mobagel, I collaborated with mobile game developers to implement a PHP-based API for a mobile game application. Additionally, I utilized Jabber (XMPP) to create a communication iOS app.
Plurk is a prominent social website in Asia, particularly in Taiwan, known for its emphasis on privacy. During my tenure, I played a pivotal role in developing the first official version of the iOS mobile app. From conceptualization to user experience design and ultimately launching on the App Store, I was involved in the entire development process. Notably, the app achieved remarkable success, reaching the #1 rank in the social category (Taiwan) within its first week of release.
GPA 3.85/4.0
Collaborated with a professor to develop a web application using jQuery and the Google Maps API for the Illinois State Museum. The purpose of the application was to suggest popular destinations in Illinois based on user preferences and current location. Additionally, while navigating, the application would display notable landmarks and attractions in the vicinity.
I co-founded a programming club with a senior classmate, focusing on sharing knowledge and delivering lectures to a group of 20 members. Our sessions covered iOS development, as well as the implementation of algorithms and data structures using C. Additionally, I led a team that participated in prestigious programming competitions such as the ACM-ICPC (International Collegiate Programming Contest) and ACM-NCPU (National Contest for Private Universities), earning multiple honorable mentions. Furthermore, I collaborated with a graduate student to develop the first official iOS app for our campus intranet services, catering to the needs of both students and staff members.
Provenance is a cutting-edge innovation platform for AI and data, offering a comprehensive dataset marketplace and an array of end-to-end model training services. These services encompass supervised, semi-supervised, generative, and more, all accessible with just a few clicks on the user-friendly website. The platform's backend is skillfully implemented using Kotlin Spring Boot and Python FastAPI, while the frontend boasts a seamless user interface powered by Vue.js. It is fully integrated with Google Cloud Platform (GCP) for cloud resource management, lakeFS for dataset management, Stripe for secure payment processing, Sendbird for efficient messaging, and DocuSign for streamlined eSigning of agreements. Adopting a microservice architecture, the platform's services are containerized using Docker, allowing for scalability and modularity. Additionally, message queue technology ensures efficient processing of training tasks, ensuring a smooth and responsive user experience.
This project focuses on deduplication within our face extraction vector, ensuring that the same user cannot enroll multiple times. As the data source expands, the deduplication process can become complex. Therefore, a more efficient solution is imperative. Implemented in Python FastAPI and leveraging Miluvs, an open-source vector database, the service handles deduplication seamlessly. Even when comparing millions of records, results are obtained within a second. This optimized approach ensures accurate and swift identification, enhancing the reliability of our face extraction system.
Licman is a robust internal license usage tracking system that meticulously monitors daily service usage for all clients and consolidates the data into comprehensive daily reports delivered to both the business team and clients. Moreover, it boasts an alarm feature capable of detecting and alerting for any unusual activities. The service's backend is skillfully implemented using Python Flask, while the frontend is thoughtfully crafted with Vue.js, ensuring a seamless and efficient user experience. Licman efficiently manages license usage and empowers businesses to make data-driven decisions while providing clients with valuable insights.
Karma Zone is an automated strike zone system for official baseball games. It uses Python, Firebase, OpenCV, and Deep Learning to detect strikes and record high frame rate videos from multiple angles. This comprehensive system provides additional game information such as trigger time, batting angle, ball speed, and ball axis.
News Interview By ELTA.TVKarma Cam, inspired by Karma Zone, is a personal training assistant system. It utilizes multiple cameras from various angles to capture and analyze every movement of players, including swing speed, angle, pitch velocity, and displacement. The product is developed using Python and PyQt.
Artomotion websiteGooroo app is designed to help students find the perfect tutor to meet their needs. It offers features such as session booking, tutor communication, learning profile reviews, and easy course payment, all within a single app. The app is developed using Swift for the frontend and Python with GraphQL for the backend. Payment processing is handled through Stripe integration.
App StoreGooroo For Tutors is a comprehensive app designed specifically for Gooroo tutors. It empowers tutors to manage their schedules, communicate with students, access earnings, leave reviews or suggestions, and maintain their profiles, all through a single app. The iOS app is developed using Objective-C, while the Android app is developed using Kotlin with the navigation component. The backend is implemented in Python with GraphQL, and payment processing is seamlessly integrated with Stripe.
App Store Google PlayIt's a cooperation project with a leading online book retailer in Taiwan. The project involved developing an advanced recommendation engine that offered end-to-end production data-driven services to enhance their commodity designs. The engine leveraged machine learning pipelines to process NLP, image, and tabular data with imbalanced labels. To ensure efficiency and scalability, the engine was containerized using Docker, employing caret, dplyr, data.table, and xgboost in R, as well as OpenCV, numpy, and scipy in Python. The engine's API interface was established through Plumber. Notably, the F-1 score achieved an impressive 0.8, showcasing the effectiveness and accuracy of the solution.
Ultron, a recommendation engine that provides personalized item recommendations to customers prior to placing their orders. The core algorithm was developed in collaboration with a data scientist , utilizing the concept of the Multi-armed bandit problem. The backend of the engine is implemented using Python Flask, with MySQL as the database and Redis for caching. The entire system is containerized using Docker. Even during peak hours at the airport, the engine effortlessly handles hundreds or thousands of requests per second on a single machine.
Concierge is the flagship product of OTG, offering a seamless hospitality experience at airports. Customers can conveniently order food, engage in gaming activities, and access airline membership bonuses, all while waiting for their flights. The app primarily leverages Objective-C for development. However, due to the integration of various JavaScript components, Concierge also incorporates WKWebView to interact with JavaScript, enabling enhanced functionality through WebViews.
News on New York TimesPlurkr, the first official Plurk app launched in 2011, achieved remarkable success by reaching the #1 rank in the social category (Taiwan) within its first week of release. The app was skillfully implemented using Objective-C, following a hybrid approach that combined WebView for certain components and native UI for the rest. This strategic choice allowed seamless styling of Plurk threads UI without necessitating app updates. Additionally, Plurkr's authentication was effectively implemented through OAuth, ensuring a secure and smooth user experience.
Plurk websiteiFCU is an iOS app that optimizes access to campus intranet services for students and professors. It offers user-friendly features like course enrollment, email access, and appointment scheduling for students. Professors can conveniently check their schedules, access emails, and review students' learning profiles. Implemented in Objective-C and utilizing SQLite, iFCU enhances campus communication and productivity, providing a seamless and efficient experience for all users.
App Store